Does Acinetobacter baumannii Serve as a Source for bla NDM Dissemination into Enterobacteriaceae in Hospitalized Patients?

Introduction

Since the beginning of the millennia, carbapenemase-producing Enterobacteriaceae (CPE) have disseminated rapidly due to the emergence of specific types of carbapenemases in different parts of the world, including the KPC, VIM, New Delhi metallo-beta-lactamase (NDM), and OXA-48 enzymes. In Israel, the CPE epidemic was initially caused by the KPC-producing Klebsiella pneumoniae, which is thought to have been introduced from the United States in 2005. ¹ The spread of this outbreak in Israel has been contained by an ongoing nationally coordinated stringent control effort. ¹ However, in recent years, several reports have indicated that other types of CPE, primarily NDM- and OXA-48-producing Enterobacteriaceae, are emerging in Israel^2,3 and are now responsible for an increasing number of CPE cases.

The mechanisms of the emergence and spread of NDM-producing Enterobacteriaceae (NDME) in Israel have remained elusive. In Israel and in other parts of the industrialized world, the first cases of NDME infections were reported in patients returning from the Indian subcontinent. ⁴ However, as soon as 2010, NDME were detected in patients without any connection with India ⁵ and now appear in various hospitals in Israel, including the post-acute care system. ³

One possible explanation for the emergence and spread of NDME in Israel, despite stringent control measures, is that there has been repeated exchange of the bla_NDM gene between Acinetobacter baumannii and Enterobacteriaceae as previously suggested. ⁶ NDM-producing A. baumannii (NDMAb) was discovered in Israel as early as 2009. ⁷ In both NDME and NDMAb, the bla_NDM gene is located within various types of transposons, most commonly the Tn125, ⁴ and can be found on both the chromosome and on plasmids.^8,9 Hence, both due to its plasmid location and from an evolutionary perspective, the almost simultaneous emergence of the bla_NDM gene in two bacterial populations is most likely explained by horizontal gene transfer (HGT) between them. ⁶ However, such events have never been traced epidemiologically nor have they been reported in other antimicrobial resistance genes.

The primary goal of this study was to study the possibility that A. baumannii serve as a source for transmission of the bla_NDM gene to Enterobacteriaceae by HGT that is epidemiologically traceable and significant. For that we performed a comparative survey of the bla_NDM gene in these two bacterial families, A. baumannii and Enterobacteriaceae. The secondary goals were to study the prevalence and molecular mechanisms of clonal- and HGT-mediated transmission of the bla_NDM gene within each bacterial family.

Materials and Methods

Results

During the 9 months studied, 13 patients carrying NDMEs were hospitalized. During this period, 313 patients had CRAB, of which 16 cases (5.1%) were NDMAb. Co-carriage of NDME and NDMAb was not found. The microbiological and molecular features of the NDME and NDMAb isolates are presented in Table 1. Most NDMAb (12/16) were isolated from clinical sites, including sputum (n = 5), blood (n = 4), urine (n = 3), peritoneal fluid (n = 1), and wound (n = 1), exclusively or combined with rectal surveillance culture. In contrast, only 5/13 of NDME were isolated from clinical sites, including wound (n = 3), sputum, and urine (n = 1 for each).

The majority of NDMAb isolates (10/16) belonged to a single PFGE type that was responsible for five clonal TEs. In all PFGE types except one the bla_NDM gene was plasmid borne and was typed as bla_OXA-66. The latter PFGE type (two isolates) was typed as bla_OXA-70 and also carried the bla_NDM-2 allele; it is likely that in these isolates the bla_NDM gene was located on the chromosome. Both E. coli and K. pneumoniae NDME belonged to diverse clonal structure. Although 9/13 cases were hospital acquired, only one clonal TE of K. pneumoniae NDME was identified. Hence, the source for most NDME cases (8/13) remained undetermined.

Despite the plasmid location of the bla_NDM gene in all NDME and 4/5 of NDMAb PFGE types, the other molecular features excluded the possibility of HGT-mediated transmission between NDME and from NDMAb in most of the cases: (1) the bla_NDM gene allele differs in E. coli (bla_NDM-5) versus K. pneumoniae NDME and NDMAb (bla_NDM-1); (2) the Tn125 was present in all NDMAb but in only 3/9 of K. pneumoniae NDME.

Epidemiologically, only one of the three Tn125-positive K. pneumoniae NDME was hospital acquired (case 375) and was also related to one potential HGT-mediated TEs from NDMAb (case 813). However, although the bla_NDM gene was located inside a Tn125 in both isolates, the extent of the similarity (75.7%) between the two transposons (Supplementary Fig. S1; Supplementary Data are available online at www.liebertpub.com/mdr) excluded the possibility of TEs. In accordance with this finding, the plasmid Inc types were also different: 375-IncFII, A4; 813-GR3.

Discussion

In Israel and in our institution, CRAB has emerged at the beginning of the millennium as an important nosocomial pathogen and a major cause of bacteremia. ¹⁹ Our study provides several novel findings regarding the epidemiology of NDMAb. First, the rate of NDMAb among CRAB was found to be relatively low (5.1%), despite the evidence for clonal spread. Epidemiological data are lacking in studies from European countries, ²⁰ but this rate (5.1%) is higher than the rates reported from Turkey (2/68, 2.9%), ²¹ the United Arab Emirates (2/155, 1.2%), ²² and China (4/2,109, 0.2%). ²³ This emphasizes the relative importance of NDMAb in Israel compared with other countries. Second, we observed an inverse proportion of clinical versus rectal surveillance culture isolates of NDMAb versus NDME. This was caused most likely by the fact that CRAB may colonize other anatomic sites that were not sampled. ¹³ Hence, the actual number of colonized patients is likely to be higher. Third, in all NDMAb isolates the bla_NDM gene was located inside a Tn125 transposon, but unlike the previous report from Israel, ⁷ the allele was bla_NDM-1 (and not bla_NDM-2), and the gene was harbored by plasmids. Together, these molecular and epidemiological features suggest that although NDMAb by itself does not constitute a major problem within the CRAB epidemic, its prevalence is similar to NDME and it possesses the potential to serve as reservoir for HGT-mediated spread of the bla_NDM gene. ⁶

From epidemiological and infection control perspectives, the lack of HGT-mediated TEs of the bla_NDM gene between bacterial families in our study is important. Although our study is too small to exclude that possibility from a broad evolutionary perspective, it points that unlike interspecies intrafamily transmission,^24,25 such events are probably too rare to be accounted as a possible source for transmission in the course of an outbreak. Therefore, identification of the bla_NDM gene should not be actively sought in CRAB-infected/colonized patients in the context of NDME outbreak.

NDM and other antimicrobial resistance genes can disseminate in the healthcare system by two molecular mechanisms. Although clonal spread is more easily noticeable, HGT of carbapenemase genes is of no lesser importance and has been responsible for the spread of these genes across states and even continents. ²⁶ Epidemiologically, HGT-mediated TEs can be traced when different CPE species harboring the same carbapenemase gene are isolated in the same patient ¹⁶ or their isolation in several patients in the context of a local outbreak. In this study we have setup a novel approach to trace HGT-mediated transmission of the bla_NDM gene between bacteria of different families. Although we have failed to identify such TEs, we believe that from an evolutionary perspective, it is very likely that such events have occurred and their study is of crucial importance for the understanding of the evolution of antimicrobial resistance. Therefore, we think that a study of such events should be sought in a multicenter study of longer duration.